(The Wall Street Journal, Nicole Nguyen, April 21, 2026)
Anthropic announced the new AI model called Mythos not for its ability to generate realistic videos or solve complex mathematical problems, but for having identified thousands of “high severity” vulnerabilities in all major operating systems and web browsers. The company decided not to release it on the open market to avoid serious consequences on economies and national security, and instead granted early access to forty large companies such as Apple, Google, and Amazon so they can identify and fix the flaws before cybercriminals exploit them. According to experts, it doesn’t matter if Mythos is truly a superpower weapon for hackers: if it’s not this model, it will be another one soon. The UK AI Security Institute has verified that Mythos can autonomously exploit vulnerabilities, performing tasks that would take a human days.
In the coming months and in 2027, companies will launch a flood of security updates and patches, with possible service disruptions and temporary system outages. Users cannot do much against these “update bottlenecks,” but they have the duty not to postpone installing those that arrive on their devices. Delaying patches gives malicious actors time to study and exploit them. It is therefore essential to enable automatic updates on smartphones, computers, and browsers, and replace devices that have passed their end-of-support date, because after that deadline they no longer receive security fixes.
To further protect themselves, it is essential to adopt good digital hygiene practices: use a password manager to create and store long, unique, and complex passwords; enable two-factor or multi-factor authentication; and switch, where possible, to passkeys, which are even more secure because they are tied to fingerprint or face and work only on legitimate sites. Finally, great caution is needed regarding AI-enhanced social engineering scams, such as vocal deepfakes of relatives in distress: creating a shared “code word” within the family can help expose these attempts. The message is clear: the future of AI will make the digital world more vulnerable, so updating immediately and strengthening personal defenses is no longer an option, but an urgent necessity.
Mythos has discovered thousands of critical vulnerabilities.
“Mythos identified thousands of ‘high severity vulnerabilities’ in ‘every major operating system and web browser.’ Anthropic feared that releasing it on the open market could have serious consequences for economies, national security, and more, so it gave early access to forty large companies including Apple, Google, and Amazon so they can fix the flaws before hackers exploit them.”
AI will make attacks more frequent and severe.
“‘Whether Mythos is truly a superpower weapon for hackers is irrelevant. If it’s not this model, it will be another one in five minutes,’ explains Dave Lewis of 1Password. The UK AI Security Institute discovered that the model can autonomously exploit vulnerabilities, performing tasks that would take a human days.”
Never postpone updates.
“You will probably see patches regularly appear as updates on your most used devices. Don’t postpone them. The longer you wait, the more time the ‘bad guys’ have to reverse-engineer the patch and figure out where the flaw is. And once they do, you can be hacked.”
Replace obsolete devices.
“Almost every gadget now has an end-of-life date. After that date, companies stop providing new features, bug fixes, and yes, critical security updates. If it’s out of support, unfortunately you will need new devices.”
Strengthen passwords and authentication.
“Improve your passwords with a password manager that creates and saves long, unique, and practically unreadable passwords. Enable two-factor or multi-factor authentication. Passkeys are even more secure: they are stored in the password manager, respond to face or fingerprint scans, replace both the password and multifactor code, and work only on legitimate sites.”
(Excerpt from the newsletter by Giuseppe Liturri)
